Child abuse and child neglect are symtoms of abnormal caretaker-child interaction. We hypothesize that such abnormal interaction often has roots in early mother-infant interaction. We, furthermore, hypothesize that certain children exhibit characteristics at birth which result in their being prone to abuse. One such category is premature infants who are overrepresented among cases of abused children. The proposed research has three major objectives: (a) to study the development and internal dynamics of disturbed interaction of mothers with infants that are at high risk of being abused (prematures), (b) to develop objective methods for the early detection of disturbed mother- infant interaction, and (c) to develop programs aimed at the prevention and/or correction of dHsturbed interaction. Interactions of mothers with full-term and of mothers with premature infants will be systematically observed during feeding. Mother and infant behaviors will be recorded through the use of mutually exclusive and exhaustive codes when the infant is three days, two weeks, six weeks, and three and one- half months old. In addition, infants will be examined separately at each of these times. These normative data will be analyzed to identify abnormal interactions and describe infant behaviors which correlate with such abnormal interactions. Specific patterns and sequences of behavior that characterize both normal and abnormal interactions will be described. Intervention programs consisting of infant stimulation alone, mother training alone, and both will be developed for premature infants on the basis of the normative data. The relative effectiveness of these programs to improve mother-infant interaction will be evaluated. It is hoped that this research will (a) result in a better understanding of the etiology of child abuse and child neglect and (b) provide effective methods for their prevention.